ATM-mediated ELL phosphorylation enhances its self-association through increased EAF1 interaction and inhibits global transcription during genotoxic stress

Abstract

Mammalian cells immediately inhibit transcription upon exposure to genotoxic stress to avoid fatal collision between ongoing transcription and newly recruited DNA repair machineries to protect genomic integrity. However, mechanisms of this early transcriptional inhibition are poorly understood. In this study, we decipher a novel role of human EAF1, a positive regulator of ELL-dependent RNA Polymerase II-mediated transcription in vitro, in regulation of temporal inhibition of transcription during genotoxic stress. Our results show that, besides Super Elongation Complex (SEC) and Little Elongation Complex (LEC), human ELL (aka ELL1) also forms a complex with EAF1 alone. Interestingly, contrary to the in vitro studies, EAF1 inhibits ELL-dependent RNA polymerase II-mediated transcription of diverse target genes. Mechanistically, we show that intrinsic self-association property of ELL leads to its reduced interaction with other SEC components. EAF1 enhances ELL self-association and thus reduces its interaction with other SEC components leading to transcriptional inhibition. Physiologically, we show that upon exposure to genotoxic stress, ATM-mediated ELL phosphorylation-dependent enhanced EAF1 association results in reduced ELL interaction with other SEC components that lead to global transcriptional inhibition. Thus, we describe an important mechanism of dynamic transcriptional regulation during genotoxic stress involving post-translational modification of a key elongation factor.

Figure 1.

ELL is required for basal level expression of multiple SEC target genes within mammalian cells. (A) Immunoblot analysis showing interaction between ectopically-expressed FLAG-ELL and different endogenous SEC components. (B) Immunoprecipitation of endogenous ELL showing its interaction with other SEC components within mammalian cells. (C) qRT-PCR analyses showing significant downregulation of expression of multiple SEC target genes upon knockdown of ELL as compared to control scramble cells. The relative RNA expressions were analyzed by normalizing the target mRNA expressions with that of actin (as internal control). (D) ChIP analyses showing reduced recruitment of SEC components on target genes at the promoter proximal region upon ELL knockdown. IgG was used as control in our experiments for calculating the fold enrichment. (E) qRT-PCR analyses showing significant downregulation of expression of multiple SEC target genes upon knockdown of AF9 as compared to control scramble cells. The relative RNA expressions were analyzed by normalizing the target mRNA expressions with that of Actin (as internal control).

Figure 2.

EAF1 negatively regulates expression of native ELL target genes within mammalian cells. (A) qRT-PCR analyses showing negative effect of EAF1 overexpression on expression of native ELL target genes within 293T cells. 293T cells were transfected with plasmid constructs expressing empty vector (EV) and FLAG-EAF1 respectively. The relative RNA analyses were analyzed by normalizing the target mRNA expressions with that of Actin (as internal control). The inset panels show overexpression of EAF1 at protein level. (B) The left panel represents ChIP analyses showing negative effect of EAF1 overexpression on recruitment of different SEC factors on ELL target genes at the promoter proximal regions. The fold enrichment of different factors on target region had been normalized with that of IgG control. The top right panel shows relative amount of Pol II at the indicated coding regions of target CCND1 and c-MYC genes. The bottom right panel shows pausing index of Pol II (a ratio of Pol II at promoter proximal region/coding region) at the target CCND1 and c-MYC genes. (C) Immunoblot analysis showing enhanced interaction between ELL and other SEC components upon EAF1 knockdown. (D) qRT-PCR analyses showing enhanced expression of different ELL native target genes in EAF1 knockdown cells. The relative RNA analyses were analyzed by normalizing the target mRNA expressions with that of Actin (as internal control). (E) ChIP analyses showing increased recruitment of different SEC components on ELL target genes at the promoter proximal regions upon EAF1 knockdown. The fold enrichment of different factors on target region had been normalized with that of IgG control. (F) qRT-PCR analyses showing reduced expression of several snRNA target genes upon EAF1 knockdown within mammalian cells. The relative RNA analyses were analyzed by normalizing the target snRNA expressions with that of Actin (as internal control). For Figures 1 and 2, all of our qRT-PCR analyses for ChIP and RNA analyses, the error bar represents mean ± SD and statistical analyses were performed using one/two tailed Student's t test wherein * denotes P ≤ 0.05, ** denotes P ≤ 0.01, *** denotes P ≤ 0.001, and ns denotes ‘not significant’. Data represents a minimum of n = 2 biological replicates and three PCR replicates for each sample.

Figure 3.

EAF1-containing ELL complex does not associate with other common SEC components. (A) Immunoblot analysis showing the absence of EAF1 in ELL•CDK9•SEC in 293T cells. The left panel shows the experimental strategy for tandem affinity purification used for this experiment and the right panel represents the immunoblot analyses to identify the interacting proteins using factor-specific antibodies in each step. (B) Immunoblot analysis showing the absence of EAF1 in ELL•CDK9•SEC through reconstitution of protein complex using baculovirus-mediated expression of indicated target recombinant proteins in heterologous Sf9 cells. The left panel shows the experimental strategy for tandem affinity purification used for this assay and right panel represents the immunoblot analyses to identify the interacting proteins using factor-specific antibodies in each step. (C) Immunoblot analysis showing formation of a separate ELL•EAF1 complex other than ELL•SEC through reconstitution of protein complex using baculovirus-mediated expression of indicated target recombinant proteins in heterologous Sf9 cells. The left panel shows the experimental strategy for tandem affinity purification used for this assay and right panel represents the immunoblot analyses to identify the interacting proteins using factor-specific antibodies in each step. (D) Immunoblot analysis showing the absence of EAF1 in ELL•AF9•SEC through reconstitution of protein complex using baculovirus-mediated expression of indicated target recombinant proteins in heterologous Sf9 cells. The left panel shows the experimental strategy for tandem affinity purification used for this assay and right panel represents the immunoblot analyses to identify the interacting proteins using factor-specific antibodies in each step. (E) Glycerol gradient-based separation of protein complex present in the nuclear extract of 293T cells. Nuclear extract was loaded onto a 4–20% glycerol gradient and was separated by centrifugation. Fractions were collected and individual fractions were tested for presence of indicated proteins by western blotting.

Figure 4.

ELL is a self-associated protein and self-associated ELL does not contain other SEC components. (A) Immunoblot analysis showing self-association between HA- and FLAG-tagged ELL within mammalian cells. (B) Immunoblot analysis showing self-association between purified recombinant ELL proteins at physiological salt concentration (150mM NaCl) in vitro. (C) Native-PAGE analyses and subsequent coomassie staining showing formation of multimeric ELL complexes by purified recombinant ELL protein. Presence of monomer, dimer and trimer species are indicated by •, ••, and ••• respectively. (D) Glutaraldehyde cross-linking based assay showing presence of monomer, dimer and higher oligomeric species of ELL within whole cell lysate of mammalian 293T cells. (E) Immunoblot analysis showing self-association ability of different ELL domains within mammalian cells. (F) Immunoblot analysis showing defective self-association in vitro between ELL molecules upon deletion of N-terminal 44 amino acids. (G) Immunoblot analysis showing the presence of EAF1 with self-associated ELL that does not interact with other SEC components. The left panel shows the experimental strategy for tandem affinity purification used for this experiment and the right panel represents the immunoblot analyses to identify the interacting proteins using factor-specific antibodies in each step. (H) Immunoblot analysis showing markedly reduced presence of self-associated ELL in ELL•CDK9 complex in 293T cells by tandem affinity purification strategy.

Figure 5.

EAF1 enhances ELL self-association both in vitro and in vivo within mammalian cells. (A) Immunoblot analysis showing presence of EAF1 in the oligomeric (dimer and above) species of ELL, but not in monomer. (B) Immunoblot analysis showing enhanced ELL self-association upon co-expression of EAF1 within mammalian cells. (C) Glutaraldehyde cross-linking based assay showing presence of enhanced formation of higher oligomeric species of ELL within whole cell lysate of mammalian 293T cells upon over-expression of EAF1. (D) Immunoblot analysis showing enhanced ELL self-association in presence of EAF1, using purified recombinant proteins in vitro. (E) Immunoblot analysis showing reduced self-association among ELL proteins upon EAF1 knockdown in 293T cells. (F) Immunoblot analysis showing the abilities of ELL domains to interact with EAF1 and other SEC components. (G) Immunoblot analysis showing the ELL domain (45–621)that is defective of EAF1 interaction, shows enhanced association with other SEC components. (H) Immunoblot analysis showing failure of EAF1 protein to enhance ELL self-association between full-length (1–621) and N-terminal deletion fragment (45–621) in in vitro self-association assay. (I) Immunoblot analysis showing reduced ELL self-association upon overexpression of EAF2 within mammalian cells.

Figure 6.

ATM-mediated ELL phosphorylation enhances EAF1 association and its self-association upon genotoxic stress. (A) Immunoblot analysis showing an increased self-association between ELL proteins in response to genotoxic stress within mammalian cells. Post 36 h of transfection with indicated plasmids, 293T cells were treated with doxorubicin for 2 hrs and the lysates were subjected to anti-FLAG immunoprecipitation followed by western blotting analysis using factor-specific antibodies as indicated. (B) Immunoblot analysis showing increased ELL self-association, EAF1 interaction and concomitant reduced SEC interaction (CDK9 as representative) within mammalian cells upon exposure to genotoxic stress. (C) Immunoblot analysis showing increased EAF1 and reduced SEC component (CDK9 as representative) association with endogenous ELL within mammalian cells upon doxorubicin treatment. (D) Immunoblot analysis showing defective genotoxic stress-induced ELL self-association in EAF1 knockdown cells. (E) Immunoblot analysis showing increased ATM-mediated phosphorylation of ectopically-expressed ELL within mammalian cells upon doxorubicin treatment. (F) Immunoblot analysis showing increased ATM-mediated phosphorylation of ELL and concomitant enhanced interaction with EAF1. (G) Immunoblot analysis showing genotoxic stress-induced increased ATM-mediated phosphorylation of endogenous ELL and concomitant enhanced EAF1 interaction within mammalian cells. (H) Immunoblot analysis showing presence of key ATM target sites at the C-terminus of ELL (521–621) which undergo phosphorylation in response to genotoxic stress. (I) Immunoblot analyses showing defective stress-induced ATM-mediated phosphorylation and EAF1 interaction with ELL triple mutant (TM, T551A, S561A, S589A). (J) Immunoblot analysis showing defective stress-induced self-association ability of ELL TM.

Figure 7.

Self-association-defective ELL mutants fail to inhibit global transcription upon genotoxic stress and thus increases DNA damage and reduces cell survival. (A) Immunoblot analysis showing reduced ability of ELL (45–621) mutant, when compared to full-length, to interact with EAF1 and concomitant increase in other SEC component association (P-TEFb complex as representative) within mammalian cells upon exposure to genotoxic stress. (B) Nascent RNA transcription analysis showing enhanced global transcriptional activity in cells expressing ELL (45–621) mutant, when compared to full-length ELL upon exposure to genotoxic stress.293T cells were transfected with plasmids as indicated. 36hrs post transfection, cells were treated with or without doxorubicin for 2hrs as indicated. 15min prior to harvesting, EU was added for its incorporation into nascent RNA. These cells were then processed (as mentioned in the Methods section) for identifying overall global transcriptional activity as indicated. The data, as presented, represents normalization of signals obtained from green fluorescence (representing nascent RNA) over red fluorescence (representing ELL proteins). The right panel represents the quantification of the overall observation. (C) Immunoblot analysis showing reduced ATM mediated phosphorylation of ELL TM and concomitant decreased EAF1 interaction when compared to ELL WT within mammalian cells upon exposure to genotoxic stress. (D) Nascent RNA transcription analysis showing enhanced global transcriptional activity in cells expressing ELL TM, when compared to ELL WT upon exposure to genotoxic stress. Experiments were performed following the procedure as mentioned in (B). The data, as presented, represents normalization of signals obtained from green fluorescence (representing nascent RNA) over red fluorescence (representing ELL proteins). The right panel represents the quantification of the overall observation. In our nascent RNA transcription assays error bar represents mean ± SD, representative of two independent experiments (n = 80 cells). Statistical analysis was performed using two tailed t-test wherein, **** denotes P ≤ 0.0001. (E) Immunoblot analysis showing increased level of y-H2AX in cells expressing different ELL mutants (45–621 and TM) as compared to WT ELL within mammalian cells under genotoxic stress. (F) Immunoblot analysis showing increased level of DNA damage marker y-H2AX in ELL knockdown cells. (G) Immunoblot analysis showing increased level of DNA damage marker y-H2AX in EAF1 knockdown cells. (H) Immunoblot analysis showing the effect of re-expression of wild type ELL and its mutant derivatives in ELL knock down cells on the level of DNA damage marker y-H2AX. (I) Cell proliferation assay showing reduced proliferation ability of cells expressing self-association-defective ELL mutants as compared to WT after 2 h of doxorubicin treatment.